"visualizing" multi-dimensional backtesting/optimization results?

M

mizhael

"visualizing" multi-dimensional backtesting/optimization results?


Hi all,

Lets say after waiting for a few days, I've gotten a 8-dimensional Sharpe Ratio array...

What's the best way to "visualize" or make sense out of the 8-D array?

Are there tools in R that allow us to "explore" this multi-dimensional array?

I am thinking of the following "flattened" structure, and hopefully writing to CSV file and opening in Excel will help a bit?

But I am not totally sure... and I am really newbie on this...

The flattened structure can be described in the following forms:

count=0
for (i1 in Parameter1Set)
for (i2 in Parameter2Set)
for (i3 in Parameter3Set)
for (i4 in Parameter4Set)
for (i5 in Parameter5Set)
for (i6 in Parameter6Set)
for (i7 in Parameter7Set)
for (i8 in Parameter8Set)
{
count=count+1
results[count, 1]=i1
results[count, 2]=i2
results[count, 3]=i3
results[count, 4]=i4
results[count, 5]=i5
results[count, 6]=i6
results[count, 7]=i7
results[count, 8]=i8
mySharpeRatio=RunBacktest(i1, i2, i3, i4, i5, i6, i7, i8)
results[count, 9]=mySharpeRatio
}

Any suggestions/thoughts would be greatly appreciated!

Thanks a lot!
 
Make contour plots -- you only need 28 of them. The human eye+brain is very good at interpreting contour plots.

Math: binomial coefficient C(8,2) = 8*7/2 = 28.
 
Quote from MGJ:

Make contour plots -- you only need 28 of them. The human eye+brain is very good at interpreting contour plots.

Math: binomial coefficient C(8,2) = 8*7/2 = 28.
More...

Thanks! I have searched a lot of websites and also received some feedbacks to my post, but it seems there is no better way than doing the pair-wise 2D visualization pair-by-pair...
 
Quote from Rodney King:

Yeesh, now OP cross-posts to the R list(s) http://bit.ly/uQoNJ7... he's "the gift that keeps on giving."
More...

And now also cross-posts to nuclearphynance -- without mentioning his cross-posting at any of the venues. Scummy, scummy, scummy behavior. :eek:
 
You must log in or register to reply here.
Back
Top